Industrial IO retrofit

Projects General Guidance
1

Hi All,

I just want to run a project by you to see if you can spot any obvious flaws or suggest alternatives. It's a simple enough project, but I am doing it for someone else. I'm spending their money on the parts, so I want to increase the chance of me getting it right first time!

I'm helping a guy with some old industrial equipment containing a failed PLC. The equipment is so old there's no chance of getting a direct replacement, or support from the manufacturer. Additionally PLCs seem to have a high entry cost, which I can't justify for a one-off project.

So my plan is to replace the PLC with an Arduino. The downside is the system currently have around 30x 24V inputs, and 16x 110V outputs. Everything on/off, no analog. Outputs are all inductive loads; motors, contactors, solenoids etc. Inputs are generally buttons, but a couple of higher frequency signals (eg motor RPM) - still quite slow, about 100rpm.

So plan is; Arduino mega with a terminal shield. 24V to 5V opto isolation for all inputs. 5V relays for all outputs.

Sound ok so far?

The bit where it gets interesting is finding the right bits. Ideally to save me work I'd like some PCBs that are pre-built & available in the UK (so shields/dev boards rather than discrete components). I can only find up to 8-channel opto boards on eBay/Bangood type places. I'd need 4 of those. Then a relay module, 16 channel, again super cheap Chinese.

I have two issues here - the first is that once I add all the wires it'll be a mess, but I guess I can cope. The second is that this will be used daily - I'm concerned that the cheapest relays in existence wont be up to the task. I could make my own board or use discrete DIN mounted relays (but finding them in 5V is a challenge - and finding somewhere for the flyback protection even harder) or replace the relay board relays with branded ones.... I don't know, the more I think about it the more work I'm creating for myself & it gets out of hand.

What do you think? Would you do differently?

Thanks!

2

Good to use opto for inputs. Keeps noice away.
It looks like 24 volt is commonly used. I suggest 24 volt powered relays on driver boards equipped eith fly back diodes. Make sure the relays handles the current needed. Arduino only handle 20 mA per pin.

3

a PLC is often a simple backplane with a power module and a CPU module then add-on's for the I/O's

the more modern PLC's have more built in, but the common brands have simple backplanes.

if the old CPU failed, and the I/O's are good, you have industrial I/O's that are not only rated for the machine, but are known to work on that machine. I suspect that there are no 20 amp power relays in the PLC blocks, that the internal relays are closer to 5 amp rated. if power relays or motor starters are used, your relay power is less than the motor needs.

#1) get a large metal enclosure with back plate. these are common for industrial work Hoffman is best known name.
#2) create a wiring diagram label all circuits and use wire markers to label the wires to match.
#3) I would opt for DIN rail mounted relays in rows
check our relays for availablily, some have manual buttons, some have internal LED's
#4) lots of colors of wires 300 or 600 volt rated (Dont' use plenum rated) inside of your box.
This may be overkill, but when in a more industrial environment, the ratings reduce questions and problems.
#5) draw a line in the box. everything on this side is connected to the outside world, everything on that side is DC power.
if space allows a DIN rail of screw terminal blocks offers a barrier strip to allow field wires to end at that point.
then, the wires go from that point into the box to the relays. size wire to the rated loads.

google PLC wiring click images. if you have room, use wiring duct. it comes in lots of sizes.
you will see that often the logic is on one side, power on the other.

a bag(s) of hair clips the ones that can grasp your little finger. helps keep wires under control.
get wire strippers. for 16-18ga wire, I like the ones that are larger and squeeze and strip in one motion and have a stop so the exposed wire is all the same length.

if you are getting new pushbuttons, use the industrial form factor, use the 22mm for smaller buttons and indicators to save space. get an actual hole punch for them. might seem expensive, but it worth it for a permanent project.

if you go with relay modules, use 7 of 8, leave one unused. make sure any wires for any others can reach that relay.
if one fails, it offers fast fix.

You can laser print wire names. take packing tape over to make a plastic surface and secure on the panel, then place a relay. makes trouble shooting easier.

as a note, each Arduino has a maximum current per channel, but a full board max current that is often much lower than the sum of all channels.

4

amphour:
I'm helping a guy with some old industrial equipment containing a failed PLC. The equipment is so old there's no chance of getting a direct replacement, or support from the manufacturer. Additionally PLCs seem to have a high entry cost, which I can't justify for a one-off project.

In full consultation with your client you need to consider very carefully who is going to maintain this system in the future taking account of the possibility that you will not be available - perhaps due to holidays, or illness or because you have moved to live somewhere else.

A lot of the price of a PLC is due to the fact that they are familiar to technicians and the client can get easily get support when there is a problem or if he wishes to expand the system.

The savings from a cheap Arduino solution would evaporate very quickly if a production system is halted for a few days.

...R

5

How well is the control system that the PLC implemented understood?

I wouldn't be surprised if there are some safety aspects in it that the users are ignorant of (because they never saw them) or have forgotten.

6

amphour:
What do you think? Would you do differently?

Have you seen these?

IMO, it'd be easier to translate the original ladder diagram to another ladder rather than to C++. And, think about down the road. Will you be available a year from now for maintenance/upgrading? If you're not and this is needed it would be much easier for a PLC programmer to come in and do it in a familiar language and not have to wade through C++.

Yes, industrial parts are more $$$ than hobby stuff but consider that they come with some support, spares, testing, certifications, etc. behind them.

YMMV

7

Go with a PLC. It was designed for this. With all that I/O to handle, it may even be cheaper in the end, as well as more dependable.

8

The PLC stuff is hardened for industrial environments, I had Allen Bradley PLC's that ran for 10 years with nobody doing any work on the systems. The A/B stuff should have the ability to swap the controller out by just sliding it out of the rack. If the processor is way old and no longer made there should be the ability to translate the Ladder Logic to the newer processor.

You are sounding like redoing the entire system, not a trivial project. the labor costs will be many times the cost of the hardware. I would stay with a PLC.

9

Just curious, what brand/model of PLC?

10

Robin2:
In full consultation with your client you need to consider very carefully who is going to maintain this system in the future taking account of the possibility that you will not be available - perhaps due to holidays, or illness or because you have moved to live somewhere else.

A lot of the price of a PLC is due to the fact that they are familiar to technicians and the client can get easily get support when there is a problem or if he wishes to expand the system.

The savings from a cheap Arduino solution would evaporate very quickly if a production system is halted for a few days.

...R

Really spot on. The money saved by hobby stuff will be lost, and more, "the day support" is needed, backup of changed parameters, restore after a crash, You name it.
I was asked by my dentist once to copy my home made and home used auto kalendar system. I said no just because every serious business must have solid product, bakup handling, service, maintanence....
The first job, after my exam, was in ADH, Automatic Data Handling. Being first man on the job every morning, answering customers asking for immedeate help most every day......

11

Fantastic responses, thanks all. It's given me a lot to think about. I'll print this thread out and take it with me next time I see him.

I should clarify that it's to get a friend out of trouble rather than a customer. They have tried the official route but the original supplier will not support it. They got quotes for someone replacing the PLC, but as suggested it's the labour costs that are high.

If I can get another couple of years use out of it then it gives them a bit more time to save up for a new one (well, they could already buy a like-for-like replacement, but they really want a much better more expensive one!). I do however agree with you all that the PLC is the 'right' solution, but less familiar to me. I'll discuss it with him for sure.

dougp:
Just curious, what brand/model of PLC?

LG Master-K series, I don't have the exact model to hand. The machine itself is around 30 years old. Any PLC documentation is gone even with the manufacturer (they said they only support for 10 years, obviously they want to sell a new one!) so translating the original ladder to anything else will not be possible. I can't even reuse the program memory chip - it is smeared along the forks of the forklift that smashed into it. A direct hit! Thanks for the link to the Automation Direct PLCs. There are some well priced options there. Hopefully I can find them in the UK.

wildbill:
How well is the control system that the PLC implemented understood?

I wouldn't be surprised if there are some safety aspects in it that the users are ignorant of (because they never saw them) or have forgotten.

I tested all IO devices and created a flow diagram of how it's supposed to work and compared it with the users understanding of the functions. From a software standpoint (I write software/firmware as a job) I am confident I can replicate what they have. Unlike my day job however, I normally have a team of people to assist with the hardware! You are dead right with the safety aspects - it's not just all the interlocks they don't know about, but the ones they've deliberately sabotaged too! Me: "The finger guard switch doesn't seem to be giving a reading" Operator: "Yeah, it's on the shelf behind you".

Thanks again!

12

amphour:
I should clarify that it's to get a friend out of trouble rather than a customer.

That can make things more complicated - harder to say NO, when it is the right thing to do.

I got involved in something like this years ago (when I was younger and less experienced). A client wanted a computerised data collection system for a production line but the company was in a poor financial state and the manager was trying to keep costs to a minimum. I was not the project leader but I was creating both hardware and software. I was doing this part-time and one thing I had not been conscious of was the expectation that the "client" had of the support I would be able to give. This was something that was never actually spoken about, but it should have been. In reality, I was prepared to create the thing but I had no interest in, or available time to provide operational support. In retrospect, it was good that the project fizzled out before it was completed - and now I can't remember why it fizzled out.

It's one thing to develop a project on a part-time (or helping-out-a-friend) basis but the friend may be innocently assuming that the project will be 100% reliable with no teething troubles and may also be assuming you will fix anything at a moment's notice. That's a big burden for you to take on and could well mean the death-knell for the friendship.

This problem of unstated expectations and serious annoyance when they are not met is very common when the person commissioning the project is ignorant (in the nicest possible sense of the word) of the technology involved. And it's not just a problem at the one-man-business level - it has resulted in $millions being wasted on many huge IT projects.

...R

13

If you do decide to go with a PLC plcs.net can be a useful site.